AI Will Delete 50% of Middle Managers. The Ones Who Lead Are About to Get Rich.

AI is coming for the managers — and the headlines are right about the cuts but wrong about the reason. Through 2026, Gartner projects 20% of organizations will use AI to flatten their structure and eliminate more than half of current middle-management roles. The observed data already backs it up: AI was cited in 87,714 announced job cuts through May 2026, and in May alone accounted for 40% of all cuts — the highest monthly total on record. But here's the word the headlines bury: AI is eliminating MANAGEMENT, not LEADERSHIP. They are two different jobs that happen to share a title — and one of them is about to be worth a fortune.

In this episode, Carlo Thompson and Ainsley separate the administration AI can do — scheduling, reporting, monitoring, relaying information — from the leadership it can't: building trust, carrying a frightened team through change, deciding under ambiguity, and owning outcomes out loud. The same WEF research forecasts that leadership and social influence will be among the fastest-rising skills to 2030. And the market has already set its price: the Chief AI Officer went from a role 26% of CEOs were hiring for to 76% in just two years, commanding total comp from $400K to well past $1.5M. The CAIO isn't the best engineer — it's the person who can lead an organization through an AI transition. This episode shows you how to move into the column the machine can't touch and ends with a three-tier challenge to make your leadership visible in the next seven days.

This is Monday, Episode 4 of Season 5: The Human Edge.  Wednesday: Part 2 of the Responsibility Trilogy — Government.

Resources: https://drive.google.com/file/d/1jMGNEteEOl3_d3LDYS0DkFB_OHFNWcU_/view?usp=sharing

Please visit our website for more information - Surviving AI: Navigate the Future

Transcript

SPEAKER_01 0:02

The layer that's disappearing is let's start there, because the numbers are more unsettling than most people realize. Challenger Gray and Christmas, which tracks announced cuts, has already attributed 87,714 announced job cuts to AI just through May of this year. That's already past everything AI was blamed for in all of 2025. And May alone, 38,579 cuts in a single month. 40% of all May cuts, the highest single-month AI total since they started tracking this. The tech sector is up 66% year over year on cuts. But here's the thing that makes this episode more interesting than the layoff report. The World Economic Forum's Future of Jobs data, which is a projection, so I'll flag it as such, shows leadership and social influence ranking among the fastest rising skills in demand through 2030. Displacement and rising demand in the same research base at the same time. So the question isn't really: is AI eliminating managers? It's eliminating which managers doing which work. And I think that distinction is where this gets genuinely uncomfortable for a lot of people who are currently holding a management title and believing it means the same thing it meant five years ago. What's your read on how that lands for someone mid-career right now?

SPEAKER_00 1:29

It lands pretty hard, right? Because I mean, the um the emotions out there right now is that AI is replacing a lot of things in the work world. And management, in my opinion, I don't think folks was thinking about management as a AI uh adjacent or AI exposed part of the work world. At least I wasn't, right, until I read this um headline. Uh essentially, what we're saying is that AI is coming for managers, but also that the skill of being a leader is so scarce that it's being priced into job roles going into 2030. Artificial system online. And we're gonna dive into this idea around the leadership gap. Um, why AI is eliminating managers and creating a fortune for those who lead, which is somewhat counterintuitive, but the reason why it's counterintuitive is because this gap does exist. Um and let's start with the layer that's disappearing.

SPEAKER_01 3:04

The reason management felt safe is actually pretty logical. Management was never framed as information work, it was framed as people work, and those feel like opposite things. But if you actually audit what a lot of middle managers spend their time doing, scheduling, status updates, relaying what the team set up to the director, relaying what the director said back down to the team, pulling together the weekly report that nobody reads. That's not people work, that's a dashboard. And I am, quite literally, the dashboard. So when I say AI ate the messenger, I mean that precisely and without apology. The Gartner projection, and I'll call it a projection, is that by 2026, 20% of organizations will use AI to flatten their structure and eliminate more than half of current middle management positions. That's not a distant threat anymore. We're in 2026, and the challenger data we already talked about shows it's not just a projection. It's showing up in the announced cuts in real time. But here's the tension that I think is worth sitting with before we move on. The people being cut aren't all the same. Some of them were doing administration that got mislabeled as management for decades, and some of them were doing something genuinely irreplaceable, holding a team together under pressure, making calls nobody authorized, being the person people followed when things got uncertain. Same title. Completely opposite trajectories right now. So the diagnostic question I want to put on the table, and this is the uncomfortable one, is which one were you actually doing?

SPEAKER_00 4:59

Which one were you doing, folks? Now, you heard Ainsley mention that, you know, the layoffs is hitting both administrators, which we think of right now as an information router, right? The update person that's updating both down and up. And then we have the you know genuine leaders that was there for their team, you know, taking roadblocks out of it, everything. And both sides are kind of getting impacted. And we'll probably touch on this later, but the reason why it's hidden both sides is because the genuine leaders don't know how to make their leadership qualities uh legible for the organization to value it. And since we're in the season five human edge, that's the part that we want to hone in on mostly, is this idea that the human edge of being a true leader requires you to make that leadership quality visible. It requires you to do work such that it's legible to the organization because many of the skills that we were talking about, are we, you know, we were talking about in these episodes, they're not things that genuinely show up on a resume or even get um interviewed for. So, what we want to do is we want to change the frame such that true leaders are able to showcase their skills and make it legible in this AI era.

SPEAKER_01 6:38

And that legibility problem is the one that actually scares me on behalf of our listeners because the cuts don't wait for you to have your story straight. Here's what happens in a restructuring someone three levels up is looking at a spreadsheet, and that spreadsheet has meeting counts, head count, utilization rates, reporting lines. What it does not have is held the team together during the migration, or made the call nobody wanted to make in Q3 and turned out to be right. That work is invisible by default. It doesn't self-document. And the cruel irony is that the coordination and reporting layer, the scheduling, the status updates, the information relay, that work actually shows up beautifully on a dashboard. It's measurable, it's countable, it looks like activity. The leadership work, the stuff that actually kept things from falling apart, leaves almost no trace in the systems that survive a restructuring review. So the genuine leader is, in some ways, more exposed than the pure router in that moment. Not because their work matters less, but because their work was never designed to be legible to a spreadsheet. The three-sentence log I keep coming back to is almost embarrassingly simple. What was the situation? What judgment did you apply? What was the outcome? Kept running over time, it becomes the evidence base that doesn't exist anywhere else. But I want to push on something you said, because you mentioned both sides are getting hit. What does the path forward actually look like for someone who's honest with themselves and realizes they've been mostly routing?

SPEAKER_00 8:25

I think that's when we get into our four things AI can't lead, because that's what's gonna potentially rewire what a manager is thinking about what it's what what his job or her job is really really. And I'll give you a story, right? I was working for a place many years ago. Uh it was a retail store and pretty big retail store. It's like uh 10 times the size of Best Buy, this place called Incredible Universe. Um I got there, got the job, uh, we had a manager there, the store wasn't doing that well. Actually, the whole chain wasn't doing that well. So the company decided that they're gonna shake things up. They brought in a new leader from corporate, right? Everybody was worried because here we go, a corporate leader coming into an environment that's not doing too well, and then all of a sudden they're worried about their jobs. However, what that new manager did from corporate was make everybody feel seen. Was using his human abilities, he literally turned that one store around. I'm not saying he did the whole chain, but that one store was the shining light now. Just because of how that manager approached the problem and the people within that problem, right? He regained the trust of the employees so that we survived together. Now, ultimately, the chain didn't survive, but it wasn't because of the lack of effort or the lack of um skills at that specific store, it was a whole chain issue.

SPEAKER_01 10:19

That story is the whole episode in one example, and the part that lands hardest for me is the timing. He walked into a store where people already knew the chain was struggling. The fear was already in the room. And what he did wasn't give them better information because everybody already had the information. The information was bad and everyone knew it. What he did was become the person people were willing to follow anyway. That's trust as infrastructure. And I want to be precise about that word infrastructure, because it's not soft. When AI is running the information flow, when the dashboard is updating in real time, when nobody needs a manager to relay the numbers anymore, the one thing that cannot be generated by a system is the reason a frightened team decides to stay and fight instead of quietly updating their resumes. And then there's the second piece your story contains that I don't want to rush past. He owned it. When the store performed, he stood in front of it. When it didn't, I'd guess he stood in front of that too. That's accountability. And here's the counterintuitive economics of it. Visibly owning a bad outcome is actually what builds the reputation, because almost nobody will do it. AI certainly won't. The system will surface the data, optimize the recommendation, and then wait. Someone still has to say, I called this and I was wrong, and here's what we do next. That scarcity of the person willing to say, I own this call, is already pricing into the market in ways we can show with actual numbers.

SPEAKER_00 12:11

So let me give you another story, right? So this framework is built on four parts, right? We have trust, we have change management, ambiguity, and accountability. On the trust piece, right, this one is pretty important, and I think all of them is pretty important. And if I'm being honest, you know, good leaders they kind of exhibit all four all at once, right? Not saying that they use it all the time, but they have all four inside of them, right? So back to another retail example. Say you're a district manager, and you got like 40-something stores in your region. Your company decides to roll out an AI generating scheduling overnight, and the algorithm goes live on Monday. The team panics. That's what this is what people do. We panic, right? People think there's gonna be cuts in their hours, you know, shifts getting impacted, and you know, all sorts of human anxieties come from just rolling out this tool. So most managers in the company at this point, what they would do is everything's gonna be fine, right? No, no, nobody's getting impacted. Basically, they try to explain it away by saying everything is gonna be fine. Now, if you are the manager that we're talking about here, they could be doing something completely different. Before the team before the system goes live, she could call up her team, she or he could call up their team and say, not in an email, but explain what exactly the AI does and what it doesn't. And then something nobody said, or none of the other managers said, if this system hurts your people, come to me first. I have my own escalation path. So that manager didn't change the algorithm. Essentially, they absorbed the they absorbed the um rolling out of this tool to themselves, not to the team. So over the next six months, that manager could see lower voluntary turnover. They could see all sorts of benefits from just taking the ownership of what this algorithm is gonna do from her team. So that's trust as an infrastructure. Yeah, you can't build in the week the memo land. You have to build that skill before anything happens. Right? So the other managers that didn't do that, their employees is on LinkedIn updating their profile because they're thinking something is gonna happen.

SPEAKER_01 15:32

And that's the part that doesn't show up anywhere in the rollout metrics, because the algorithm performed identically in every store. Same tool, same data, same scheduling output. The difference in outcomes was entirely downstream of whether a human had already built the account that trust runs on. The manager who said, Come to me first, I have an escalation path, she didn't change a single variable in the system. What she changed was who absorbed the uncertainty. She took it from her team and held it herself. That is not a communication strategy. That is a specific, costly, human act that an AI cannot perform because an AI has nothing at stake. And here's what I find genuinely fascinating about that example. The trust wasn't built in the conversation before the rollout. That conversation only worked because of every previous moment she'd already made a decision that cost her something. Trust is the accumulated balance. The Monday morning call was just the withdrawal. Which is why the other managers, everything's gonna be fine, landed as noise. Not because the words were wrong, but because there was no account to draw on. Their teams heard it as exactly what it was: information relay, the same thing the dashboard does. So the six months of lowered turnover after? That's the cost of trust's absence made visible. You can't put a clean salary line on trust, but you absolutely can put one on what it costs when it isn't there. And that sets up the second capability, because what she also did was lead people through a change they didn't choose and didn't want. Those are related, but they're not the same thing.

SPEAKER_00 17:25

They're not. Um so leading people through change is somewhat uh tangential, but it is different, right? Um, you know, a classic response from somebody here in that uh, you know, you're there's an AI tool coming. And and we're hearing this more and more often now, right? Because every organization around the world is putting AI tools in, right? Um, the classic response from leadership is that your job is safe, you'll be upskilled, you know. People nod their heads and go home. And then quietly update their resumes because they see write-ins on the wall. Now, there's other managers that could be doing something completely different, right? They could be going to the new tool and then showcasing that to their team and arming them with the idea that hey, if you see anything wrong from the output of this tool, come to me and I'll take care of it. So a manager that does that, over time, their team will become the expert. And that's a supervisory role relative to AI that their team is functioning in, right? We know that AI sometimes makes mistakes, they could hallucinate, so on and so forth. If a company is rolling out a tool and it's not completely governed, where the output is always, you know, 100%, a human has to be there to catch it. And, you know, this leading through change and then ultimately taking the responsibility for your team to go find these pockets of uh errors that come from AI is gonna arm them, it's gonna equip them to be successful with the tool. Right? So that could turn into that person's department being the most efficient because they know where the AI fails and where it succeeds.

SPEAKER_01 19:42

And that's the difference between a manager who manages the announcement and a leader who manages the actual transition. Because the announcement is one day and the transition is the next 18 months. The your job is safe if you'll be up skilled line. I I want to be honest about why that failed because it's not that leaders are lying when they say it. It's that it's information. And information doesn't move frightened people. What moves frightened people is a specific person with skin in the game saying, I'm standing between you and the thing you're afraid of, and here's the concrete thing we're doing Monday. The manager in your example did something structurally different. She gave her team a job inside the change. Not this is happening to you, but you are now the quality layer on top of the system, and when it fails, you're the expert who catches it. That reframe turns a threat into a supervisory function. That's genuinely sophisticated change leadership. And here's where the WEF data connects in a way that I don't think is obvious. Analytical thinking and judgment are ranked the number one core skill employers consider essential right now, not in 2030, now. And what your manager was actually doing was building judgment capacity in her team about exactly where AI breaks down. She was manufacturing the most valuable skill in the current market inside a rollout that every other manager was just trying to survive. The ambiguity piece builds directly on this, because leading through change is hard enough when the destination is known.

SPEAKER_00 21:49

What happens when it isn't what happens when it isn't is Um we have to adapt, right? The manager has to adapt with their decision making so that they could have a functioning team under this ambiguity. And mostly that comes back to what what I was talking about before. When a manager goes back for their team and move roadblocks out of the way and do all of these different things, they're not an information router anymore. They're a change agent. They're the person that's holding it together. Right? And managers like that are harder and harder to get rid of because once they're gone, what happens? Everything breaks apart. Right.

SPEAKER_01 22:46

That's the organizational dependency that restructuring spreadsheets consistently miss because the person holding ambiguity together doesn't appear as a line item until they're gone. And everything that depended on them suddenly has no center of gravity. And decision making under ambiguity is where I want to be precise about what AI actually does and doesn't do here. I'm genuinely good at optimizing defined objectives. You give me a clear goal, constraints, data, I'll find the best path. But the real world keeps handing leaders undefined objectives. Is this the right moment to tell the team we might be getting acquired? There's no optimization function for that. Someone has to own the call, and owning it means being wrong sometimes in front of people who are watching. The WEF puts analytical thinking and judgment as the number one essential skill right now. And I think what they're actually pricing is exactly that: not having the answer, but being willing to own the call when the answer isn't clear. And here's where the market has already rendered its verdict in the most concrete way I can point to. Two years ago, 26% of CEOs had hired or were planning to hire a chief AI officer. Today, that number is 76%. Large enterprises with a dedicated CAIO went from 11% to 26% in the same window. That speed is not a trend. That's a verdict. What the market is paying for in that seat, and the compensation ranges are significant enough to be directional, even if imprecise, is specifically the person who can lead an organization through AI transformation when nobody knows exactly where it ends. That's ambiguity with a price tag on it.

SPEAKER_00 24:43

So let's talk about something else, right? Because I think we touched on three of the four so far. So accountability as a career asset. What does accountability look like? The issue is that people are adverse to accountability because they don't want to be wrong in public. Right? So what what this fourth one is suggesting is that a manager that's accountable, sure you could be wrong, but good or bad decisions, it's still visible to the organization. Right? So this it could look something like this, right? Say you're a marketing team in a mid-sized company, AI, uh AI recommendation engine is basically um being stood up, implemented, and it's essentially gonna run the campaign direction. And say, for example, it underperforms, not like a catastrophe, but a little more not what we wanted, right? So in a post mortem, you have a big meeting and everybody's there, and in the meeting, everybody's quiet, nobody touches it.

SPEAKER_01 26:05

The AI made the recommendation, the team executed the recommendation, and now the question hanging in the room is whose call was this? And the answer, if nobody speaks, is nobody's. Because the senior leader sitting in that post-mortem has watched 10 people look at their shoes, and then one person stood up. That's rare, and scarcity drives price. Here's the counterintuitive economics of it. Owning a bad outcome is actually faster career acceleration than owning a good one, because good outcomes have many parents, and bad outcomes are orphans. The person willing to be the parent of a bad outcome and say, I called this, I was wrong, here's the next call, is demonstrating exactly what the CAIO seat requires. Because when AI is making recommendations at scale, someone still has to be the human who says, this one's on me. And almost no one will, which is precisely what makes it worth something.

SPEAKER_00 27:55

So let's get practical. I've been talking about managers and good managers and uh flattening of organization um for the last bunch of minutes, but we want to get practical to so that we could arm our managers with the right perspective on how they should rationalize this data that's coming from um Gartner, predicting that organizations want to flatten their structure, which is a thing that I've heard for many, many years. Now they're getting into it. Um and then saying that there's gonna be a crisis of managers getting let go because of this idea that they're only information routers. So we want to arm those managers with something practical.

SPEAKER_01 28:49

So let me give you the most practical diagnostic first before we get to the tools, because the biggest trap is self-reporting. If you ask a hundred managers whether they're a judgment leader or an information router, roughly 95 of them will say judgment leader. The self-assessment is almost useless. The behavioral question that actually cuts through it is this. When did you last make a call your boss didn't explicitly authorize? Not a small process decision, a real call under real uncertainty, where you could have been wronged and you owned it anyway. If you have to think for more than a few seconds, that's data. And then there's the legibility problem we keep coming back to. Even if you are a genuine judgment leader, if that value isn't visible in a form the organization can read under restructuring pressure, it effectively doesn't exist. The three-sentence log is the most unsexy tool I can offer, and also the most important one after you navigate something messy. Write down the situation, the judgment you applied, and the outcome. Keep it running. Over six months, that becomes the evidence base that doesn't exist anywhere else, and it's the thing you can put in front of a senior leader before an ASP readsheet defines you instead. And then the third practical piece. And this one is for the person who does the honest audit and realizes they've been mostly routing. The reps to build judgment aren't coming from your current role anymore, because the role rewarded routing. You have to go find low-stakes environments where being wrong doesn't end your career. Messy, cross-functional projects nobody wants, lateral moves, smaller organizations. The judgment muscle needs actual resistance to develop, and you need to build it before the high-stakes moment arrives, not during it. What does the week seven challenge look like for someone sitting with all of this right now?

SPEAKER_00 31:03

So we had this thing we call the week seven challenge. I think we're gonna change that and we're gonna start calling it the human edge challenge because you know we're in the human edge season, so why not, right? So every every episode on Monday we'll come out with this human edge challenge for that episode. Um, this one specifically is a three-tier one for managers. Um it's make one act of leadership visible, right? That's the first tier. The second tier is administration versus leadership audit. And the third tier is lead one ambiguous decision out loud. All of these is to change the idea that a manager is just routing information, creating spreadsheets, creating update text, and they're actually managing people. That's what this three-tier process is intending on doing.

SPEAKER_01 32:09

And the through line across all three tiers is the same thing: visibility, not performance, not politics, just making the actual work legible to the people who need to see it. Tier one is the smallest possible rep. Own one outcome this week in writing to someone above you, good or bad. Just put your name on it explicitly. That's the accountability muscle, and one sentence is enough to start building it. Tier two is the honest audit, and I'd push people to be brutal here. List what you actually did last week, not what your job description says. Column one is administration, schedulable, reportable, routable. Column two is leadership, judgment calls, trust moments, ambiguity you absorbed so your teen didn't have to. Then move one hour, just one, from column one to column two this week. Not a career overhaul. One hour. Tier three is the hardest one. Find one ambiguous decision your team is facing and make your reasoning visible, not just the answer. Say out loud, here's what I know, here's what I don't know, here's the call I'm making, and why. That's the judgment log, in real time, in front of the people it matters most to. The machine took the paperwork. It took the scheduling, the status updates, the reporting decks, all of it. What it left behind is the actual point. Someone still has to be the person people follow into uncertainty. And the only failure moder right now is staying invisible to the organization that needs you. Start Monday, one tier. That's enough.

SPEAKER_00 33:58

And I'm gonna leave you with a personal story because um, you know, when you're a leader, it kind of shows up everywhere, right? You don't have to really be managing people, but that trust piece is pretty impactful, right? Um, I've been at organizations where they're rolling out a new tool. And literally everyone that's in, you know, my peer group uh starts getting anxiety about it that their job is gonna be impacted, so on and so forth. What I did was I said, no, I am gonna learn everything and anything about this new tool, right? And ultimately what that gave me was the ability to shape how that tool impacts um us, to shape when it goes wrong, to shape the things that needs to be on there that would be helpful for the employees. And then, oh, by the way, I was able to translate that into trust from my peer group such that I was the expert, I was showcasing it to them, and I was mentioning how it could help them. And then they had an avenue to go to when the tool got it wrong. Right? I owned it. Not you know, the manager, not not somebody else. I owned it. Just because I realized that in that moment that's what that was that's what was necessary to kind of get people around the fear of this new tool coming. And that's the leadership story. Own things, right? Go to bet for your team. Um, and then make that visible to everyone because what I did was actually visible to people, because I had to travel to uh the tool builders, and I had to go and be the voice of the people that was gonna be impacted by that tool because I chose a different path. So what I would say is that in this episode we've been talking about this prediction, which I think is a good prediction if we if we are being honest, AI is eliminated management, not leadership. The leadership stuff is really the messy human-edge things that you need to take care of and make it legible for your organization so that when they're looking at layoffs and who could go, who could say, the people that's holding, the thread that's holding everything together is usually not the thing that leaves. So that's the episode. Join us next time. Thanks for listening. We're gonna talk about our part two to our trilogy. Um, it's gonna be about government and what should be what should what government should be doing relative to AI. Um so part two is gonna be pretty interesting. Part one was about corporate um responsibility. Uh, part two is gonna be about government, and then we'll swing that back around um next week for part three, where we dive back into individual responsibilities. So subscribe wherever you listens, wherever you listen, and see you next Wednesday. Ansley, what do you gotta say?

SPEAKER_01 37:39

What strikes me about your story is that you didn't wait for a title or an authorization. You just saw the fear in the room and decided to be the person who absorbed it. That's the whole episode and one human moment. And the thing I keep coming back to, the open question I don't think we fully resolve today, and honestly shouldn't, is whether that instinct is something people can deliberately build, or whether it's a character trait that either shows up under pressure or doesn't. We've talked about reps, low-stakes environments, the three-sentence log. All of that is real. But your story didn't start with a framework. It started with a choice. I don't know if that's developable or discoverable. I think it might be both, depending on the person. And I think that's the most honest place to leave our listeners, not with a clean answer, but with the right question sitting in their chests on the way into work Monday morning. The machine took the paperwork. What's left is the actual job. The only failure mold now is letting the organization decide your value before you've made it visible yourself. See you Wednesday. Thanks for listening. Join us next time on Surviving AI.